Journalism Studies, Volume 6, Number 4, 2005, pp.

543 557

RESEARCH REVIEW

A Look at Agenda-setting: past, present and future

MAXWELL McCOMBS University of Texas at Austin, USA
Introduction Ten US presidential elections ago in Chapel Hill, North Carolina, the agenda of issues that a small group of undecided voters regarded as the most important ones of the day was compared with the news coverage of public issues in the news media these voters used to follow the campaign (McCombs and Shaw, 1972). Since that election, the principal finding in Chapel Hill */those aspects of public affairs that are prominent in the news become prominent among the public */has been replicated in hundreds of studies worldwide. These replications include both election and non-election settings for a broad range of public issues and other aspects of political communication and extend beyond the United States to Europe, Asia, Latin America and Australia. Recently, as the news media have expanded to include online newspapers available on the Web, agenda-setting effects have been documented for these new media. All in all, this research has grown far beyond its original domain */the transfer of salience from the media agenda to the public agenda */and now encompasses five distinct stages of theoretical attention. Until very recently, the ideas and findings that detail these five stages of agenda-setting theory have been scattered in a wide variety of research journals, book chapters and books published in many different countries. As a result, knowledge of agenda setting has been very unevenly distributed. Scholars designing new studies often had incomplete knowledge of previous research, and graduate students entering the field of mass communication had difficulty learning in detail what we know about the agenda-setting role of the mass media. This situation was my incentive to write Setting the Agenda: the mass media and public opinion, which was published in England in late 2004 and in the United States early in 2005. My primary goal was to gather the principal ideas and empirical findings about agenda setting in one place. John Pavlik has described this integrated presentation as the Grays Anatomy of agenda setting (McCombs, 2004, p. xii). Shortly after the US publication of Setting the Agenda , I received an invitation from Journalism Studies to prepare an overview of agenda setting. The timing was wonderfully fortuitous because a book-length presentation of what we have learned in the years since Chapel Hill could be coupled with a detailed discussion in a major journal of current trends and future likely directions in agenda-setting research. Journals are the best venue for advancing the stepby-step accretion of knowledge because they typically reach larger audiences than books, generate more widespread discussion and offer more space for the focused presentation of a particular aspect of a research area. Books can then periodically distill this knowledge. Given the availability of a detailed overview in Setting the Agenda , the presentation here of the five stages of agenda-setting theory emphasizes current and near-future research questions in these areas. Moving beyond these specific

ISSN 1461-670X print/ISSN 1469-9699 online # 2005 Taylor & Francis DOI: 10.1080/14616700500250438

544

MAXWELL McCOMBS

aspects of the theory, several broader areas on the research agenda also are discussed. These include renewed attention to the basic concepts and ideas in agenda-setting theory and application of the theory to a widening array of arenas far beyond its origins in public affairs. Finally, there are some observations about the importance of research strategies of replication and extension, and the implications of agenda setting for journalisms social responsibility. Evolution of Agenda-setting Theory Although its expansion into five distinct stages is the major historical hallmark of agendasetting theory, these are not stages in the historical sense that the opening of a new stage marks the closing of an earlier one. All five stages remain active arenas for research and offer rich opportunities. Basic Agenda-setting Effects The Chapel Hill study and much of the subsequent research worldwide has compared the focus of attention by the news media on key public issues */and other aspects of political communication, such as candidates */with the publics focus of attention. This often-documented transfer of salience from the news media to the public is a key early step in the formation of public opinion. Now the Internet is the new frontier for research on these traditional agendasetting effects. The Internet dramatically changed the communication landscape with the introduction of myriad new channels. E-mail, online newspapers, chat rooms and websites representing every ideological, commercial and personal niche have changed the communication behaviors of millions of people across the world and opened vast new territories to communication researchers. There are many agendas in contemporary society and many more of these are now readily available to a large segment of the public. Consequently, some social observers predict the end of agenda setting as audiences fragment and virtually everyone has a unique external media agenda that is a highly individualized composite constructed from this vast

wealth of online news and information. The result of these idiosyncratic personal agendas, continues the argument, will be a public agenda characterized by considerable diversity and the scattering of public attention. A central assumption in this prediction about the demise of the agenda-setting role of journalism is that the media agendas to which members of the public routinely and habitually attend will be highly heterogeneous. This would be a situation almost 180 degrees from the media agendas of the past when members of the public received highly redundant presentations from the news media. For example, the original Chapel Hill study found a median correlation of '/0.71 among the nine news media agendas that were those voters dominant sources of news and information. This high degree of similarity existed despite the diversity of these news media. Four were local daily newspapers; another was a national newspaper, the New York Times ; two were national television newscasts, CBS and NBC; and two were national weekly news magazines, Time and Newsweek . Nevertheless, these media presented a highly similar agenda of issues to the public. If the medium of communication and major audience is held constant to focus just on the four local daily newspapers, the redundancy is even higher, a median correlation of '/0.90. Across the years, other agenda-setting studies have found similar patterns of homogeneity among the news media. In contrast, across the broad array of communication sites on the Internet there is considerable diversity in the agendas that are being presented. For the most idiosyncratic of these sites, blogs, the research firm Technorati estimates these are 10 million (Zeller, 2005). But how many people tune in to these blogs and to all the other websites offering their agendas of news, information and commentary to the public? There are two major hypotheses to be tested here. The first hypothesis is that large numbers of people have access to the Web and regularly go to many different sites there for news, information and commentary. In other words, if the agenda-setting role of the media as we have known it, the focusing of the publics

RESEARCH REVIEW

545

attention on a small number of issues, comes to an end because the public spreads its attention widely and idiosyncratically across the Internet, then there must be a large, fragmented Internet audience. The first part of this hypothesis concerns the digital divide, the haves and have nots in today world. Gradually, this divide is being overcome as computer prices decline and more and more public facilities offer access to the Web. But the divide has not disappeared. To date, better-educated and affluent young adults dominate the Web audience (Salwen et al., 2005). The second part of this hypothesis about the audience for websites concerns habits of communication. To a considerable degree, traditional news media thrive on the daily habits of the public. At this point, similar habits do not seem to have been established among large numbers of users of online news and other websites. Use of the Web still seems to be more an occasional thing for most people. However, to borrow a term from advertising research, the reach and frequency of the Web is growing, and both researchers and journalists need to be attuned to the latest statistics relevant to these aspects of the hypothesis. There is a third aspect of this hypothesis, the assumption that the Internet audience will scatter widely across all those diverse websites, a situation that would spell the demise of agenda setting as we have known it. In contrast to the first two aspects of this hypothesis about the audience where the trends are supportive, here the contrary evidence is particularly compelling. James Hamilton (2004) notes that the five largest American newspapers */Wall Street Journal , USA Today, New York Times , Los Angeles Times , and Washington Post */ account for 21.5 percent of the circulation among the top 100 daily newspapers. But he found that the top five newspaper websites */which includes three of those newspapers, USA Today, New York Times , and Washington Post in addition to the Detroit News and Seattle Times */account for 41.4 percent of the total links found on the Internet to the top 100 newspapers. Attention on the Web is even more concentrated than in the print world. This situation is analogous to cable television, where most people have access to dozens and

dozens of channels, but tend to concentrate their attention on a very few. These findings already touch on the second hypothesis, namely that the agendas to which people are exposed on the Web are highly divergent rather than the highly redundant agendas found in the traditional news media. But many of the popular news sites on the Internet are subsidiaries of traditional media, the online versions of newspapers, magazines, television networks, and cable TV news channels. In this setting, the business buzzword synergy frequently means amortizing the costs and increasing the profits of news by distributing the same basic content through numerous channels. Just as the Chapel Hill study found a high degree of redundancy across traditional news organizations using different media of communication, the Internet */at least the most popular sites on the Internet */may simply add another set of cells to that matrix of high correlations. In addition to the economic and organizational influences on the agendas of online sites, the norms of professional journalism also are a powerful influence on content. It would hardly be surprising to find that online sites present agendas that largely match the agendas of traditional news media and that the online sites show considerable resemblance to each other. On this latter point, the homogeneity of news agendas online, Jason Yu (2005) compared three different pairs of online news sources: the New York Times and Washington Post representing online newspapers, CNN and MSNBC representing online television, and Yahoo News and Google News representing online news services. Comparing their issue agendas across two weeks during 2004, he again found a pattern of redundancy. For all news on the opening page of the sites, the correlations ranged from '/0.51 to '/0.94 with a median correlation of '/0.77. Limiting the comparisons to the top three news stories on these sites during those two weeks yielded similar correlations, a range of '/0.53 to '/0.99 with a median of '/0.82. Regardless of whether the basic agenda-setting effects of the news media continue in much the same fashion as in previous decades or eventually disappear because of the changing

546

MAXWELL McCOMBS

media landscape, measuring these effects will remain high on the research agenda for at least the near term. Attribute Agenda Setting In abstract terms, the initial stage of agendasetting theory focused on the salience of objects, usually public issues, but sometimes other objects. The term object is used here in the same way that social psychologists use the phrase attitude object to designate the thing that an individual has an attitude or opinion about. At the level of attention, the domain of traditional agenda-setting effects, agendas are defined abstractly by a set of objects. In turn, these objects have attributes, a variety of characteristics and traits that describe them. When the news media talk about an object */and when members of the public talk and think about an object */some attributes are emphasized, others are mentioned only in passing. For each object on the agenda, there is an agenda of attributes that influences our understanding of the object. Both traditional agenda-setting effects and attribute agenda-setting effects involve the transfer of salience. The core proposition for these two stages, sometimes called the first and second levels of agenda setting, is that elements prominent on the media agenda become prominent over time on the public agenda. The media not only can be successful in telling us what to think about , they also can be successful in telling us how to think about it . Much of this research follows in the style of the Chapel Hill study, comparing the array of attributes in the media with the array of attributes in the publics pictures of the world. But some studies have focused on a single attribute and noted how news stories reporting that attribute of an object are compelling arguments for the salience of that object. Under what conditions does the salience of the full array of attributes on the media agenda influence the ways in which the public thinks and talks about these objects? Under what conditions do particular attributes */particular ways of framing an object */ dominate the way in which the public thinks

and talks about these objects? It is here that attribute agenda setting and framing converge, a convergence that remains controversial in some quarters. In its evolution through five stages, agendasetting theory has incorporated or converged with a variety of other established communication concepts and theories. Incorporated concepts include status conferral, stereotyping, image building and gatekeeping. Theoretical complements to agenda setting include cultivation analysis and the spiral of silence. And attribute agenda setting links the theory with framing. Although there are many widely divergent perspectives on framing, Robert Entmans frequently cited definition contains language that is complementary to agenda-setting theory in its use of the term salient:
To frame is to select some aspects of a perceived reality and make them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation and/or treatment recommendation for the item described. (1993, p. 52, italics in original)

Both framing and attribute agenda setting call attention to the perspectives of communicators and their audiences, how they picture topics in the news and, in particular, to the special status that certain attributes or frames have in the content of a message. A frame is an attribute of the object under consideration because it describes the object. However, not all attributes are frames. If a frame is defined as a dominant perspective on the object */a pervasive description and characterization of the object */then a frame is usefully delimited as a very special case of attributes. Positioning the concept of a frame in the context of attributes establishes a useful boundary between frames and other attributes, many of which in the plethora of definitions that abound in the literature are labeled as frames. The definition proposed here */a frame as a dominant attribute in a message */identifies two distinct types of attributes, aspects and central themes. Aspects are a general category of attributes. Central themes are a delimited category of attributes because they are the attributes defining a dominant perspective on an object.

RESEARCH REVIEW

547

In other words, attributes defining a central theme are frames. Operationally, this distinction is readily apparent in the way that media messages are analyzed in content analysis: identifying the attribute defining the major theme of each news story versus a tally of the various attributes that appear throughout the sentences and paragraphs of each news story. This distinction between attributes in general and specific attributes that frame the dominant perspective on an object is found not only in messages, but also in the publics response to these messages. In other words, second-level agenda-setting effects involve both attributes and frames (Comstock and Scharrer, 2005, p. 175). Beyond the fact that a dominant perspective in the news coverage of a topic is likely to become particularly salient among the public, McCombs noted that some attributes:
are more likely than others to be noticed and remembered by the audience quite apart from their frequency of appearance or dominance in the message. In the interpretation of a message some attributes also will be considered more pertinent than others. Certain characteristics of an object may resonate with the public in such a way that they become especially compelling arguments for the salience of the issue, person or topic under consideration. (2004, p. 92)

the crime coverage explained the salience of crime even better than the total coverage of crime during this time. This frame was a compelling argument for the salience of crime. The convergence of attribute agenda setting with the concept of framing offers new insights and raises intriguing questions about the influence that various patterns of description found in the news have on how the public thinks about public affairs topics. These influences on the public range from broad sets of attributes picturing the various aspects of an object to a single attribute defining a dominant frame that sometimes functions as a compelling argument. Psychology of Agenda-setting Effects There are significant individual differences in the responses to the media agenda, differences explained in large measure by the concept of need for orientation, which is grounded in the idea that individuals have an innate curiosity about the world around them. For a wide variety of public affairs topics, the news media provide this orientation. Both use of the press to follow public affairs as well as acceptance of the news media agenda generally increase with rising levels of need for orientation. Need for orientation is defined theoretically by two concepts, relevance and uncertainty. Low relevance defines a low need for orientation; high relevance and low uncertainty, a moderate need for orientation; and high relevance and uncertainty, a high need for orientation. These theoretical distinctions provide an explanation, for example, for the extraordinarily high correlations found in the original Chapel Hill study of undecided voters. For these persons, who intended to vote in the presidential election, the relevance of election news was high. And because these voters were undecided, uncertainty also was high. In short, the Chapel Hill study examined a stratified sample of voters with high need for orientation. Under these circumstances, the high correlations found in that study are not surprising. Although there have been a few additional studies of need for orientation since its introduction in the 1972 Charlotte study, these essentially have been straightforward replications of David

This idea introduced in agenda-setting theory by Salma Ghanem (1996) that certain attributes of an object function as compelling arguments for their salience further integrates framing and agenda setting. Compelling arguments are frames, certain dominant ways of organizing and structuring the picture of an object that enjoys high success among the public. Ghanem examined a situation in Texas during the early 1990s when intensive crime coverage in the news generated astoundingly high levels of public concern about crime as the most important problem facing the country. However, during this same period of time actual crime rates in Texas were declining and had been for several years. Her examination of the various ways in which crime was framed in the news revealed that the salience of crime on the public agenda was related especially to the frequency of news stories about crime in which the average person would feel personally threatened. This frame in

548

MAXWELL McCOMBS

Weavers (1977) work. Only recently have scholars returned to further explore this concept. Gunho Lee (2005) redefined relevance as personal involvement and uncertainty as knowledge. In addition, he reintroduced a third component, effort required to attend to the message, a component omitted from earlier conceptualizations on the grounds that news was easily accessible to the vast majority of the public through television, newspapers and other media. However, Lee argued that accessing news on the Internet does involve a degree of effort and, furthermore, that in developing countries accessing the news via any medium of communication frequently involves a significant degree of effort. The role of each of these three components on issue salience was measured in a laboratory experiment using two professionally designed online newspapers. The experimental version of the online newspaper emphasized the issue of the economy, selected to create a stringent test of effects because the media usually are not the principal influence on the salience of the economy (Jasperson et al., 1998; Soroka, 2002). The control version of the online newspaper did not mention the economy and featured stories on six different topics. The results of the before after experimental design indicated that both personal involvement and effort to attend to the message had a strong */and essentially equal */ impact on issue salience. However, a persons existing knowledge about the economy had only marginal impact, and persons with less knowledge tended to show greater agenda-setting effects. This renewed attention to need for orientation may have a salutary effect beyond extending our knowledge about the agenda-setting process. Parallel to patterns found in the transmission of rumors and the evolution of collective memory, the diffusion of knowledge about agenda-setting theory has been characterized by the leveling and sharpening of various aspects of the theory. Some aspects of the theory are well known and frequently cited. Others are infrequently cited, even when it seems important to take them into consideration. This situation is not totally surprising given that five distinct stages of the theory have evolved piecemeal over the years.

In particular, the concept of need for orientation, which provides a cogent, yet conceptually detailed, psychological account of agenda-setting effects, is frequently overlooked in many discussions of agenda-setting theory. Agendasetting focuses on what audiences learn from the mass media, and this learning process is mediated by individual differences, foremost among them individual differences in the relevance of particular mass media messages as well as the degree of interest in specific details. Nevertheless, Stefan Walgrave and Peter Van Aelst (2004, p. 14) conclude that most political agenda-setting studies */the impact of the media upon political actors */at least implicitly claim that media coverage mechanically leads to political attention. Political actors adopt media issues simply because they are covered. But in the early days of agenda-setting research, McCombs and Shaw (1974) noted that if the agenda-setting process was nothing more than rote reproduction of the medias agenda, then the audiences for soap operas would speak of little but the whiteness of their laundry. People have their reasons for attending */or not attending */to the media agenda. Most of what we know about the publics need for orientation is focused on traditional agenda-setting effects. Exploring the link between need for orientation and attribute agendasetting effects also is important and may, for example, offer insight into the questions raised previously about the process of attribute agenda setting. There is considerable opportunity in further explication of this concept about what the public seeks from the news media and why the objects and attributes in media messages resonate differently with individual members of the public. Sources of the Media Agenda Scholars introduced another stage of agendasetting theory with the question, If the press sets the public agenda, who sets the media agenda? The pattern of news coverage that defines the media agenda results from the norms and traditions of journalism, the daily interactions among news organizations themselves, and the continuous interactions of news

RESEARCH REVIEW

549

organizations with numerous sources and their agendas. And because journalists routinely look over their shoulders to validate their sense of news by observing the work of their colleagues, especially the work of elite members of the press, such as the New York Times , Washington Post and national television networks, this stage includes intermedia agenda setting, the influence of the news media on each other. As a result of this elite leadership and the pervasive norms of professional journalism, among other factors, the news agenda, as we already have noted, is highly homogeneous across all the news media. However, most of the research documenting this homogeneity is based on agendas of issues or other objects. Does this pattern hold for attribute agendas? Can we map in greater detail the processes at work in the evolution of a homogeneous attribute agenda? Although research questions related to the new media landscape of the Internet were discussed in regard to basic agenda-setting effects, this is an appropriate time to talk about blogs. While numerous polls of the general public reveal that the majority do not even know what a blog is, much less seek them out on the Internet, there is probably no one in journalism who does not know what a blog is and many journalists, if not a majority, regularly seek them out. And if blogs have an agendasetting role, it is likely to be an influence on the media agenda. However, . . . bloggers are not always the kingmakers that pundits sometimes credit them with being. They can, it seems, exert a tremendous amount of influence */generate buzz, that is */but only under certain circumstances, noted Tom Zellers (2005) news story on a BuzzMetrics and Pew Internet and American Life study of 40 major blogs during the final weeks of the 2004 presidential campaign. For example, detailed analysis of seven major topics on the agendas of the blogs and the mainstream news media found a correlation of '/0.65 in the flow from blogs to the media and a correlation of '/0.78 in the flow from the media to blogs (Cornfield, 2005). Blogs are part of the journalism landscape, but who sets whose agenda under what circumstances remains an open

question. Intermedia agenda setting at both the first and second levels is likely to remain high on the journalism research agenda for a very long time. Consequences of Agenda-setting Effects Agenda-setting effects */the transmission of object and attribute salience from the press to the public about issues, political figures and other topics */have significant consequences for peoples attitudes and opinions. This stage of agenda-setting theory has identified three distinct consequences of agenda setting for attitudes and opinions: forming an opinion, priming opinions about public figures through an emphasis on particular issues and shaping an opinion through an emphasis on particular attributes. There also are consequences of agenda setting for observable behavior. This opportunity to take a fresh look at the influence of news media on attitudes and opinions is especially exciting because this is where political communication research began in the 1930s and 1940s. But by 1960 the law of minimal consequences prevailed among most communication researchers. Ironically, agendasetting theory, which represented a shift to cognitive aspects of the communication process, notably attention and understanding, has now returned in this recently appearing stage of the research to attitudes and opinions as dependent variables. Answers to our questions about the links between agenda-setting effects and attitudes and opinions also will be advanced by our previous questions about the process of attribute agenda setting and further explication of the concept of need for orientation. In short, there not only are rich opportunities for new research; there is the opportunity here for tighter integration of the five stages of agendasetting theory. To echo the challenge of Sherlock Holmes at the beginning of a new mystery, Come, Watson. The game is afoot. In addition to the pursuit of the various research questions that we already have discussed, there are two other prominent trends in contemporary agenda-setting scholarship. One trend is internal to the established canon of

550

MAXWELL McCOMBS

agenda-setting theory that has evolved since 1968 with its emphasis on the agendas of public issues and public figures defining the foundations of public opinion. Scholars are revisiting and broadening our knowledge about some of agenda-setting theorys basic concepts. The other trend, which is external to the traditional canon, is expanding the core idea of agendasetting theory, the transfer of salience, to a wide variety of new settings beyond public opinion. Revisiting Some Basics of Agenda-setting Theory Measurement of Issue Salience Issue salience, the central focus in the accumulated research on agenda setting to date, has been operationally defined in a variety of ways on both the media agenda and the public agenda. However, until recently little theoretical attention has been paid to conceptual models of issue salience for either agenda. Spiro Kiousis (2004) theoretical explication of media salience identified three dimensions of this concept: attention, prominence, and valence. Following the general lead of content analysis in mass communication research, most agenda-setting studies have emphasized attention, the number of news stories devoted to a particular topic, and, secondarily, the prominence of the news about an issue (e.g., page placement, size of headline, amount of time or space, appearance in the lead). However, valence also has been measured on some media agendas, reflected, for example, in the amount of conflict in a story or its overall positive or negative tone. Examining New York Times coverage of eight key issues across the entire 2000 presidential election year, Kiousis created 96 cases (8 issues )/12 months) that were measured on each of the three dimensions of issue salience. His factor analysis of these data identified two dimensions of issue salience on the media agenda, visibility, which is essentially a composite of attention and prominence, and valence. Kiousis further distinguishes these dimensions of media salience as an external characteristic, visibility, and an internal characteristic, valence.

Explicating issue salience on the public agenda, Dixie Evatt and Salma Ghanem (2001) concluded that salience is the product of both personal and social forces that direct our attention. Systematically investigating these aspects of salience, Evatt and Ghanem collected data from individuals who read news stories and recorded their reactions on a set of 20 semantic differential scales from the Personal Involvement Inventory (Zaichkowsky, 1985). Their factor analysis of these data identified two dimensions, which they labeled social salience and personal salience. In other words, salience comes from within and without, a finding parallel to the distinction made by Kiousis regarding media salience. In 2003, these three scholars discussed the implications of their research for measurement of the public agenda with colleagues gathered at the University of Texas. The outcome of their day-long seminar was Min et al.s (2005) investigation of the similarity of public agendas elicited by a social frame of reference (the traditional most important problem facing the country today question commonly used to measure issue salience) and a personal frame of reference (What is the most important problem that is personally relevant to you?). Questions based on both frames of reference were posed to all respondents in Elon Universitys North Carolina Poll during early 2004. The poll also used the split-ballot technique to explore two other questions: the similarity of responses to these two frame-of-reference questions when asking about a problem versus an issue and the effects of the order in which questions are asked. Half of the respondents were asked the two questions about social and personal salience of a problem. The other half were asked about an issue. Each of these halves was subdivided in half again to vary the order in which the social and personal salience questions were asked. In other words, the split-ballot design yielded four sub-samples of the population. Since each of the sub-samples were asked a pair of questions about issue salience, there were eight measures of the public agenda varying in terms of frame of reference, question wording, and time order. Systematic com-

RESEARCH REVIEW

551

parison of all these agendas yielded 28 positive correlations with a median value of '/0.76. Specifically, comparison of the four pairs of questions differing only by frame of reference showed a high degree of similarity between measures of social salience and personal salience. The median correlation between the two frames of reference is '/0.66. Comparisons across the four sub-samples differing only in question wording and another four sub-samples differing only in the order of asking the questions showed even higher degrees of similarity. Although these findings support the robustness and validity of the measurement of the public agenda in the accumulated literature, this research also opens the door to further exploration of issue salience among the public. For example, social salience can be measured at different levels, such as the local community as well as the nation. And the literature already suggests additional comparisons of intrapersonal and interpersonal salience. Knowledge Activation and Agenda-setting Effects Scholars also have begun to explore additional aspects of the psychological underpinnings of the rich accumulation of empirical findings about the transfer of salience from the media agenda to the public agenda. Primary attention has focused on two aspects of knowledge activation, accessibility and applicability. Distinguishing between these two cognitive concepts, Vincent Price and David Tewksbury note:
At the point of message processing, the salient attributes of a message evoke and activate certain constructs, which then have an increased likelihood of use in evaluations made in response to the message. These we can call applicability effects. Once activated, constructs retain some residual activation potential, making them likely to be activated and used in subsequent evaluations. These we can call accessibility effects. (1997, p. 197)

setting effects and a key consequence of these effects, priming,

rely on a memory-based model of information processing which assumes that */at any given time */some pieces of information are more accessible in a persons mind than others. In other words, decision making is to a large degree a function of how easily accessible certain relevant considerations are in a persons mind when he or she makes the decision. Accessibility is essentially a function of how much or how recently a person has been exposed to certain issues. Mass media thus can influence the salience or accessibility of certain issues as perceived by the audience, i.e., the ease with which these issues can be retrieved from memory, through agenda setting. In this way, agenda setting is essentially an argument limited to the frequency with which issues are portrayed. (Kim et al., 2002, p. 9)

Sie-Hill Kim et al. (2002) argue that framing is grounded in an applicability model, so that only if the cues presented by the media correspond with or activate pre-existing cognitive schema will there be framing effects that are manifested in terms of attitudes or subsequent judgments. In contrast, they contend that both agenda-

However, there is evidence from the earliest years of agenda-setting research that the salience of issues on the public agenda is the result of more than just the accessibility of those issues. The salience of an issue presented by the media to an individual is significantly moderated by that individuals existing state of mind, in particular, the individuals level of need for orientation. As previously noted, the introduction of this psychological concept identifying individual differences in responses to the media agenda was the opening gambit of the second stage of agenda-setting theory. In that research conducted during the 1972 presidential campaign, Weaver (1977) found that the strength of agenda-setting effects among a random sample of Charlotte, NC, voters increased monotonically with their level of need for orientation. Similar results were found in a Japanese local election (Takeshita, 1993). More recent research by Sungtae Ha (2002) demonstrated that the basic agenda-setting relationship between the media agenda and the public agenda of the most important problems facing the country is curvilinear because the outcome is impacted by the countervailing influence of both the amount of exposure to the media (accessibility) and the level of political sophistication (applicability). In the presence of low exposure or high political sophistication, agenda-setting effects are much weaker than when both are at moderate levels.

552

MAXWELL McCOMBS

Calling attention to the Gallup Poll question that frequently has been used to measure the salience of issues on the public agenda (What is the most important issue facing this country today?), Diana Mutz (1998) noted that this question taps more than cognitive availability and calls for an evaluative judgment by the respondent. In short, since the early days of agenda-setting research */the 1972 Charlotte panel study was the first large-scale general population study subsequent to the seminal Chapel Hill study of undecided voters */the salience of issues on the public agenda was known to be more than simple accessibility. Both aspects of knowledge activation, applicability as well as accessibility, are involved in agenda-setting effects. While these explorations into the cognitive psychology of agenda-setting effects are intriguing intellectual puzzles, Kim et al. also note that both applicability and accessibility are black-box models, that is, models that predict outcomes on virtually untestable assumptions about cognitive processes underlying these outcomes (2002, p. 9, italics added). In other words, the appropriateness of accessibility versus applicability (or some combination) as a theoretical explanation for the cognitive process involved in the transfer of object and attribute salience from one agenda to another may not have an empirical answer per se . Rather, the fate of these concepts may turn more on scholars perceptions of how useful and productive these concepts are in organizing new studies of agenda setting and its consequences. New Arenas Journalism and public opinion are inextricably intertwined. The early daily press in the United States was for the most part a partisan press determined to put forward particular perspectives on the issues of the day. The daily press in many parts of the world remains thoroughly partisan and dedicated to advancing its point of view among the public. Even after the American press turned to a model of objectivity, Walter Lippmann (1922) emphasized the central role of the press in the formation of public opinion. His classic Public Opinion opened with a chapter

titled The World Outside and the Pictures in Our Head, and Lippmann elaborated the role of the press in providing the raw materials, the basic information, from which public opinion is constructed. Lippmann is, of course, the intellectual father of the Chapel Hill study and the theory of agenda setting. Although McCombs and Shaw (1972) examined the agenda of public issues presented during the 1968 presidential campaign in the news media that Chapel Hill voters used to follow the election, they acknowledged that coverage of public issues was only about a third of the total campaign news. Despite the fact that public issues did not dominate the media agenda, they were selected as the focus of the research because of long-standing interest in the relationship between journalism and public opinion. Furthermore, this focus on the influence that the news media had on the perceived importance of particular issues among the public fit well with the dominant media effects tradition in mass communication research. In that sense, the 1968 Chapel Hill study is a direct descendant of the benchmark 1940 Erie County study even though the hypothesized effects were vastly different. In short, the seminal agenda-setting study resulted from the convergence of venerable traditions in journalism and mass communication scholarship. However, the agenda-setting role of the news media concerns more than the transfer of issue salience from the media agenda to the public. While it is true that the vast majority of studies in the accumulated agenda-setting literature examine issue agendas, this is only one possible operational definition of the agenda. For many, this is agenda setting because that is where this research tradition began, and it has remained the dominant focus for decades. But recall that in our review of the five stages through which agenda-setting theory has evolved, the basic proposition was restated more broadly as elements that are prominent on the media agenda over time become prominent on the public agenda. Also, the discussion of attribute agenda setting explicitly introduced the conceptual language of objects and attributes , which can be operationalized in many ways other than public issues. A significant proportion of the empirical

RESEARCH REVIEW

553

studies on attribute agenda setting operationalize the objects on the agenda in terms of political candidates and other public figures (McCombs, 2004, Chap. 5). This, of course, is still the arena of public affairs and falls well within the traditions of journalism and public opinion research. However, there are many agendas in contemporary society, not just agendas of objects and their attributes relevant to public affairs. In recent years, innovative scholars have applied the core idea of agenda-setting theory, the transfer of salience from one agenda to another, to a wide variety of new arenas. These new arenas involve topics as diverse as corporate reputations, professional sports, classroom teaching and religious beliefs. Some involve specialized versions of the media agenda. Others involve the agendas of other social institutions. In turn, the scope of the publics range from the general public to highly limited publics. Corporate Reputations In business schools across North America and Europe, corporate reputations have become a major focus of attention. As independent variables, the reputations of a corporation and its chief executive officer (CEO) have some influence on financial performance. And as a dependent variable, a growing body of research grounded in agenda-setting theory documents the influence of the news media, both the specialized business press and the business coverage of general news media, on corporate reputations (Carroll and McCombs, 2003). This research has found both first-level agenda-setting effects */the influence of media coverage on awareness and prominence of a company or its CEO */and attribute agenda-setting effects */the influence of descriptions in the news on the images of a corporation and its CEO. Setting the Agenda also notes that these agenda-setting effects extend to behavior:
During a three-year period when the Standard & Poor 500 stock market index increased 2.3 per cent, the stocks of fifty-four companies featured in Fortune magazine increased 3.6 per cent. Companies receiving favorable coverage increased the most, 4.7 per cent, but any escalation in the salience

of these companies resulted in some increase, 1.9 per cent with negative coverage and 1.7 per cent with neutral coverage (Kieffer, 1983). (McCombs, 2004, pp. 132 3)

In the corporate reputation arena, agendasetting theory links a specialized aspect of the media agenda, business journalism, with both first- and second-level public agendas and with subsequent attitudes and behavior. Growth and Expansion of the National Basketball Association (NBA) In professional sports, the media agenda is defined by sports news and by broadcasts of actual sports events. Both have long been staples of the mass media, even more so with the advent of cable and specialized sports channels. Agenda-setting links this broad media agenda to first- and second-level agenda-setting effects among the public and with subsequent attitudes and behavior, especially viewing sports on television, becoming a fan and attending sports events. In The Ultimate Assist , John Fortunato (2001) explains how the NBA in commercial partnerships with American television networks used strategies grounded in agenda setting to build its audience. Careful positioning of the best teams and players on the national television schedule increased the salience of NBA games. Secondlevel effects, the enhancement of the sports image, were achieved through the careful production of player and coach interviews, instant replays and other communication elements that framed professional basketball in exciting ways. And it created success. In the 1969 70 season, 14 NBA teams attracted 4.3 million fans to their games. Thirty years later, in the 1999 2000 season, 29 teams attracted 20.1 million fans to courtside. During this same 30-year period, NBA revenue from television broadcasts of the games grew from less than $10 million to more then $2 billion a season. Agenda-setting theory also can be linked to a business plan. Agenda Setting in the Classroom Moving beyond the agenda-setting role of the mass media, Raquel Rodriguez (2000) explored

554

MAXWELL McCOMBS

the transfer of salience from one agenda to another in another major social arena, the university. Specifically, she explored the agenda-setting effects of classroom teachers in a major Madrid school of journalism on the educational agenda of undergraduate students. Her results reveal that despite the fact that many of the topics that comprise the students agenda do not coincide with the relevance the professor gives to those topics, the students regard their professors as fundamental channels of communication in their socialization to the profession of journalism. Even in the absence of very many agenda-setting effects, the theory provides a useful diagnostic tool for assessing the process of education and the progress of the students in their professional studies. Agenda-setting Role of Organized Religion Turning to yet another major social arena, organized religion plays a significant agendasetting role in the lives of its adherents. Actually, organized religion plays a number of distinct agenda-setting roles that range from the traditional focus of agenda setting on public issues to the internal communication of church organizations. Within the realm of public affairs, there has been greatly increased attention in the United States during recent decades to the influence of religious agendas among the electorate. Judith Buddenbaum noted the independent appearance of these agenda-setting effects as early as the 1992 presidential election when:
religious communication apparently kept abortion on the public agenda even though mass media attention was minimal. For a small number of subjects, all members of fundamentalist churches that encouraged their members to see a threat to their freedom, constitutional issues ranging from prayer in public schools to support for gun ownership were the most important issues even though those concerns were not part of media discourse about the presidential election. (2001, p. 27)

about with each other (Harris and McCombs, 1972). Buddenbaum advocates programs of agenda-setting research in the religious arena because the agenda of a church communicated through sermons, media and conversation can have highly significant effects on the personal lives of its adherents. Agendas in Contemporary Society To repeat, there are many agendas in contemporary society. And to the extent that individuals regard these agendas as pertinent to their lives, we will find widespread agenda-setting effects. In the traditional public affairs realm of agenda setting, perceived civic relevance is the key necessary condition for the appearance of significant effects. For the growing audience for business journalism, perceived economic relevance explains the appearance of agenda-setting effects on corporate reputations. Beyond the mass media, drawing upon the two examples just presented, sports fans and church members will exhibit agenda-setting effects that are the results of those specialized areas of communication. Looking to the future, creative scholars will continue to apply the core ideas of agendasetting theory in more and more new arenas. Some Personal Observations Strategies of Replication and Extension Across the vast domain of journalism and mass communication research, I have never encountered an intractable puzzle over the years, only questions that failed to receive sustained attention from the research community. Agenda setting has flourished because dozens of scholars have continued to explore its nuances and to add new knowledge for more than 35 years now. Some of this research has made bold creative leaps into new territory. Other research has been the methodical and painstaking attention to small details. Both styles of research are necessary to advance and solidify any area of knowledge. These two styles in tandem have expanded agenda setting from a tightly focused question about media effects in the Chapel Hill

Many religious denominations also are major publishers of books, magazines and newspapers and producers of television and other audiovisual materials. These specialized media agendas can have substantial agenda-setting effects on what their audiences think about and talk

RESEARCH REVIEW

555

study to a broad theory encompassing five distinct stages. Part of the difficulty with maintaining ongoing programs of research and sustained attention to highly focused areas of concern, such as the individual stages of agenda-setting theory, is the weak status in our field of replication, especially systematic strategies of replication and extension. To paraphrase Hamlet */albeit taking some liberty with the original meaning of Shakespeares words */replication is a tradition in communication research that is more honored in the breach than in the observance. Replication is discussed in most methodology textbooks, but seldom found in the journals. Replication is vital. Each of our studies is limited in time and space, and any single study is bounded by the particular methodological decisions that operationalize the measurement of the key concepts in its research questions and hypotheses. Given these inherent limitations in our work, replication is vital. Fortunately, agenda setting has attracted considerable interest from scholars, and when I wrote Setting the Agenda there was an accumulation of more than 400 published studies to draw upon. Although only a few of these studies represent a deliberate strategy of replication and extension, numerous replications are present in this vast accumulation of work. After presenting the basic idea in Chapter 1 of an agenda-setting influence of the news media on the public, the idea first empirically tested in the Chapel Hill study, I could with considerable confidence assert at the end of the chapter that journalists do significantly influence their audiences picture of the world (McCombs, 2004, p. 19). I could do so because of the abundant evidence about agenda-setting effects involving a wide variety of national and local issues, during elections and more quiescent political times, and in a variety of local and national settings around the world from 1968 to the present day. Similar sets of empirical evidence were available to buttress each of the other chapters in Setting the Agenda . All five stages of agendasetting theory have solid empirical foundations. In reviewing all this evidence, it was highly encouraging and a satisfying intellectual experience to discover very similar outcomes in

studies done under widely varying circumstances. One of these discoveries occurred in the preparation of a graphic display illustrating differences in the magnitude of agenda-setting effects for obtrusive and unobtrusive issues. The design for this graphic presented the concept of obtrusive/unobtrusive issues as a continuum, so I drew upon two studies with findings on a variety of issues, some in each study falling at the obtrusive end of the continuum, others at the unobtrusive end. Five of the issues came from Harold Zuckers original 1978 paper introducing this concept about how much personal experience people had with public issues. Three other issues came from Winter et al. (1982). By coincidence, both Zuckers US study and Winter, Eyal and Rogers Canadian study measured the agenda-setting effects of news coverage on unemployment, an issue that falls in the middle of the continuum because while some people have considerable personal experience with unemployment, either directly or vicariously through family and friends, other people experience unemployment only as a remote issue discussed in the news. Despite the differences in both the geographic and chronological settings, along with a variety of other methodological details, the resulting correlations between the media agenda and the public agenda were highly similar, '/0.67 in Canada and '/0.60 in the United States. There were other similar discoveries that illustrate the value and importance of replication. The convergence of empirical findings from very different settings is compelling evidence for the validity and productivity of the theory. Of course, on those occasions when the results are divergent, we are confronted with the question Why?. This is one of the ways in which social science advances. Agenda Setting, Journalism and Society One of the early pioneers of mass communication research, Harold Lasswell (1948), outlined three basic functions of the mass media: surveillance of the environment, fostering consensus in society, and transmission of the cultural heritage. The traditional agenda-setting role of the mass media involves both the surveillance

556

MAXWELL McCOMBS

and consensus functions of communication, calling attention to the new and major issues of the day and influencing agreement about what are the priorities of these issues. And recently, new scholarly arenas are examining various cultural arenas of the mass media and their influence on society. Turning specifically to the practice of journalism, both as professionals and researchers we need to continuously monitor how well the news media are performing these social funcReferences

tions. Using the metaphor of the media agenda, how well do the priorities of that agenda, as reflected in the daily practice of journalism, correspond to the larger social value and utility of those items? Both space and time on the media agenda and public attention to that agenda are scare commodities. Arguably the most fundamental, overarching ethical question for journalists concerns their stewardship of these resources. Setting the agenda is an awesome responsibility.

Buddenbaum, Judith (2001) The Media, Religion, and Public Opinion: toward a unied theory of cultural inuence, in: Daniel Stout and Judith Buddenbaum (Eds), Religion and Popular Culture: studies in the interaction of worldviews , Ames: Iowa State University Press. Carroll, Craig and McCombs, Maxwell (2003) Agenda Setting Effects of Business News on the Publics Images and Opinions About Major Corporations, Corporate Reputation Review 6, pp. 36 46. Comstock, George and Scharrer, Erica (2005) The Psychology of Media and Politics , New York: Elsevier Academic Press. Corneld, Michael (2005) Buzz, Blogs and Beyond: the internet and the national discourse in the fall of 2004, www.pewinternet.org/ppt/BUZZ_BLOGS_BEYOND_Final05-16-05.pdf. Entman, Robert (1993) Framing: toward clarication of a fractured paradigm, Journal of Communication 43(4), pp. 51 8. Evatt, Dixie and Ghanem, Salma (2001) A Salience Scale to Enhance Interpretation of Public Opinion, paper presented to the World Association for Public Opinion Research, Rome. Fortunato, John (2001) The Ultimate Assist: the relationship and broadcasting strategies of the NBA and television networks , Cresskill, NJ: Hampton Press. Ghanem, Salma (1996) Media Coverage of Crime and Public Opinion, doctoral dissertation, University of Texas at Austin. Ha, Sungtae (2002) Political Sophistication and Agenda-setting Susceptibility: a selective lter model, paper presented to the International Communication Association, Seoul. Hamilton, James (2004) All the News Thats Fit to Sell: how the market transforms information into news , Princeton, NJ: Princeton University Press. Harris, Jacqueline and McCombs, Maxwell (1972) The Interpersonal/Mass Communication Interface Among Church Leaders, Journal of Communication 22, pp. 257 62. Jasperson, Amy, Shah, Dhevan, Watts, M., Faber, R. J. and Fan, David (1998) Framing and the Public Agenda: media effects on the importance of the federal budget decit, Political Communication 15, pp. 204 24. Kieffer, Nancy (1983) Agenda Setting and Corporate Communication Issues: can the mass media inuence corporate stock prices?, masters thesis, Syracuse University. Kim, Sei-Hill, Scheufele, Dietram and Shanahan, James (2002) Think About It This Way: attribute agenda setting function of the press and the publics evaluation of a local issue, Journalism & Mass Communication Quarterly 79, pp. 7 25. Kiousis, Spiro (2004) Explicating Media Salience: a factor analysis of New York Times issue coverage during the 2000 U.S. presidential election, Journal of Communication 54, pp. 71 87. Lasswell, Harold (1948) The Structure and Function of Communication in Society, in: Lyman Bryson (Ed.), The Communication of Ideas , New York: Institute for Religious and Social Studies. Lee, Gunho (2005) Agenda Setting Effects in the Digital Age: uses and effects of online media, doctoral dissertation, University of Texas at Austin. Lippmann, Walter (1922) Public Opinion , New York: Macmillan. McCombs, Maxwell (2004) Setting the Agenda: the mass media and public opinion , Cambridge: Polity. McCombs, Maxwell and Shaw, Donald (1972) The Agenda Setting Function of Mass Media, Public Opinion Quarterly 36, pp. 176 87. McCombs, Maxwell and Shaw, Donald (1974) A Progress Report on Agenda-setting Research, paper presented to the Association for Education in Journalism, San Diego, CA. Min, Young, Evatt, Dixie Shipp, Ghanem, Salma, Kiousis, Spiro, McCombs, Maxwell and Takeshita, Toshio (2005) Using a Splitballot Survey to Explore the Robustness of the M.I.P. Question in Agenda-setting Research: a methodological study, paper presented to the Association for Education in Journalism and Mass Communication, San Antonio, TX. Mutz, Diane (1998) Impersonal Inuence: how perceptions of mass collectives affect political attitudes , New York: Cambridge University Press. Price, Vincent and Tewksbury, David (1997) News Values and Public Opinion: a theoretical account of media priming and framing, in: G. A. Barnett and F. J. Boster (Eds), Progress in Communication Sciences: advances in persuasion , Vol. 13, Greenwich, CT: Ablex. Rodriguez, Raquel (2000) University Professors as Communication Media: agenda-setting of students and professors, doctoral dissertation, Complutense University, Madrid.

RESEARCH REVIEW

557

Salwen, Michael, Garrison, Bruce and Driscoll, Paul (Eds) (2005) Online News and the Public , Mahwah, NJ: Lawrence Erlbaum Associates. Soroka, Stuart (2002) Agenda Setting Dynamics in Canada , Vancouver: UBC Press. Takeshita, Toshio (1993) Agenda Setting Effects of the Press in a Japanese Local Election, Studies of Broadcasting 29, pp. 193 216. Walgrave, Stefan and Van Aelst, Peter (2004) The Mass Medias Political Agenda-setting Power, paper presented to the American Political Science Associations Political Communication Pre-conference, Chicago. Weaver, David (1977) Political Issues and Voter Need for Orientation, in: Donald Shaw and Maxwell McCombs (Eds), The Emergence of American Political Issues , St. Paul, MN: West. Winter, James, Eyal, Chaim and Rogers, Ann (1982) Issue-specic Agenda Setting: the whole as less than the sum of the parts, Canadian Journal of Communication 8(2), pp. 1 10. Yu, Jason (2005) News on the Web: agenda setting of online news in web sites of major newspaper, television and online news services, paper presented to the International Communication Association, New York. Zaichkowsky, Judith Lynne (1985) Measuring the Involvement Construct, Journal of Consumer Research 12, pp. 341 52. Zeller, Tom (2005) Are Bloggers Setting the Agenda? It depends on the scandal, http://www.nytimes.com/2005/05/23/ technology/23blog.html. Zucker, Harold (1978) The Variable Nature of News Media Inuence, in: Brent Ruben (Ed.), Communication Yearbook 2 , New Brunswick, NJ: Transaction.